Wireless Access

last person joined: 20 hours ago 

Access network design for branch, remote, outdoor and campus locations with Aruba access points, and mobility controllers.
Expand all | Collapse all

Upgrading to from partitioning failure

This thread has been viewed 49 times
  • 1.  Upgrading to from partitioning failure

    Posted Oct 23, 2021 09:54 AM

    We're upgrading a couple of clusters of 7240XMs up to AOS from however they keep failing to partition the APs and give up after 30 minutes. This happens if we do it from the mobility conductor GUI or CLI. When I issue a show lc-cluster cluster01 controller details there is no evidence of any active APs being moved about. It's as if the command has got lost in the ether somewhere. The virtual Mobility conductors are already on and are supporting a cluster of 7240XM on which were already on before being put into a cluster.

    I thought it was a bit odd so brought up a new cluster of 7210s with no APs on them and tried upgrading that via the CLI using:

    lc-cluster test-cluster initiate upgrade version partition 0

    And I get the same result - it sits there on 'AP Partitioning In Progress' for 30 minutes and fails which is surprising given there are no APs to partition. The failure message is 'AP partitioning did not complete for long time'.

    I can't see anything in the release notes to say this shouldn't work nor that I need to go via or I'm sure we did an 8.6 to upgrade in the past without an issue. Google also returns nothing that matches.

    Is this something anyone else has seen and instantly knows what bit of config I've got wrong or is it one for me to talk to TAC about?

    Richard du Feu

  • 2.  RE: Upgrading to from partitioning failure

    Posted Oct 24, 2021 05:17 AM

    I've had a look back through what's changed since we last did a successful cluster upgrade. The major thing is adding 'Bridge Role' and 'AP wired port​ profile' to some AP group profiles for AP-503h ethernet ports​. I can see how this could prevent APs moving between mobility controllers and therefore stop the partitioning process although have not been able to test this in our lab as that's a single controller.

    The good news, however, is when we've moved APs over to the cluster running from we've resolved a lot of little issues where we were having reports of things being slow, coverage not being great, performance on teams being poor and various other little annoyances. The remaining issues we have seem to be resolved with the latest windows drivers.

    Richard du Feu